Feedback from trial session
I just found your comments today. I remember why I have a hard time using the discussion threads....
I appreciate the input. I love the formulae you came up with. I'm surprised by the variability in the measurements. Makes me wonder if this approach is viable. I'll put a bunch of students onto this in my Spring course and see what sort of variability they come up with along with a comparison to original measurements.
The photos are about 3,000 pixels wide. I agree that greater precision can be achieved with a larger image. If the approach has a bias, there is no reason to expect the bias would create fake patterns...rather I'd expect either a consistent bias in one direction affecting all skulls, or increasing variance that would tend to obscure real patterns. My logic is that this is a conservative approach...if we find a pattern it's probably real.
In Spring I'll get 2 things going: more skulls; more measurements. Please thank your students for all their work. If they are interested in environmental data please let me know....land use..water quality....insect diversity. I have lots: http://www.uvm.edu/~streams/
One way around the issue with notification of active threads on watched pages, is to activate LQTemail, under gadgets in my preferences. Been working consistently for me.
I can get back to this as well in the spring. I'll have a group of bio students exploring the ideas in evolution and a group of stats students learning about t-tests and ANOVAs.
Also, an idea which came from one of my students this year is whether there would be a difference in size between male and female skulls. Do you have this classification for the skulls? It would be useful, even if there tends to be no difference.
Hope you summer is going well.
I just activated LQTemail...better late than never; thanks for the tip
Re gender.....I have no idea. I get these from trappers so all I know with certainty is where the animals lived and died. There is an interesting thought though....if interested in regression with latitude; or pairwise comparisons, one could select the largest individuals from a given region...they will usually be male because the males are larger. This would decrease variance for large scale comparisons. CHeers Declan
Thanks for your analysis ideas.
Thanks Alison You are welcome of course! Thanks for the input. I used your suggestions to improve things and presented it all at the ABLE conference (Bio Lab Educators). I had the educators at the meeting make measurements from a sample of the largest NE and largest NW skulls and take it from measurement through stats (aimed at upper level college). We did on paper measurements or used ImageJ. If you print the ruler with each skull you can use the paper ruler to measure things (only for the skull that the ruler came with).
Here is the sample we worked with: http://wikieducator.org/NE_NW_skull_comparison Might be better for high school to start with a question where you already know the answer? NE>NW
I just read that coyotes from NW have shorter snouts....something for my Evolution kids in Spring to investigate. I expanded my Alaska and Texas collections so if there's a difference we'll find it
Great idea to pick out sample with a known result, that we could work on from measuring through to statistical comparison. I have two kids in both my biology and statistics classes, so I'm hopeful that they'll be interested to follow through with the whole process. I'll definitely use the selected sample. We'll be tackling this in May or so. I'll check back here with how we make out.